The Glass Bottle Defect Detection Based On Visual Measurement

With the glass bottle market and its production scale expansion,pipeline operation and testing speed of the demand,manual testing method has been unable to follow the development of the times and market demand.At present,the research and equipment in the field of glass bottle vision inspection in our country still have a very big gap in practicality,and market share compared with the foreign visual inspection research and equipment.Especially in the visual inspection of glass products in the location of defects,types of defects collection,as well as defect discrimination analysis,there are still a lot of needs to be further studied.At the same time,with the rapid development of related computer science and related hardware and software.At present,visual inspection of glass bottle defects based on digital image processing method is the more common and advanced on-line glassware detection method at home and abroad.1.The existing glass bottle detection theory and testing equipment was be retrieved and be investigated.On this basis,through the analysis of the intelligent detection system at home and abroad,the relevant theoretical algorithms in the existing domestic and foreign glass bottle detection equipment are studied in reverse,and the principle of these theories is thoroughly studied.The main development direction and the existence problems of different equipment was understand,and the research direction of this paper was determined,and a new detection method with application value and innovation was put forward.At the same time,the equipment and related tools needed in the experiment of this paper and determine the basic scheme,index and related hardware and software selection and configuration.2.In the research process,the bottle mouth defect detection and the bottle body defect detection are taken as the research direction direction.In the part of bottle mouth detection,after contrasting various positioning algorithms,a new detection method is proposed to locate the center of the bottle mouth.Based on the engineering needs,a defect detection method is designed.The proposed method is tested,and the test results show that the proposed method meets the requirements of the detection.In the part of bottle body detection,after studying and comparing other image preprocessing methods,a new template is proposed to preprocess the image,which solves the problem of the bottle image preprocessing stage.According to the detection requirements,a method of searching for defects based on the connected domain feature search is designed and classified according to the defect features.3.On the basis of the research on the defect detection of the glass bottle mouth,the program testing platform is designed according to the requirements of the study to meet the purpose of the test.After the testing platform is built,the software is designed,and the man-machine operation detection is realized by software programming.